Encryption, the next "feature" for 2.4GHz?

Problems, experiences or just something to say about RC gear? Say it here.

Why they should not start encrypting

Postby errsemr » Sat May 29, 2010 9:42 pm

Encryption would break compatibility with existing equipment. Users hate that and would see that as an opportunity to switch to another system.

Encryption will use more processing power on the receiver (rx). The rx gets warmer and uses more electricity.

If you want to break the encryption you can do so at your desktop with much more processing power than what is practical to use in a tiny rx. You will also be able to take advantage of developements in technology.

An encrypted system is harder to debug.

Encryption sometimes (depending on the implementation) increases frame length and therefore risk of frame loss.

The user gets all the drawbacks but does not benefit.

The benefit to the vendor is questionable (but that's a topic for another thread).

Are there alternatives?
1) Rather than encryptiong the actual communication the rx may send a digital signature with it's GUI, model ID, protocol ID, a random number already received from the transmitter etc. at bind time. The transmitter can check the signature at bind time and refuse binding if the signature does not match. This gets around some but not all of the above problems and may still be too easy to break.

2) If you introduce an inovative new feature in your protocol you could patent it. This may already have been done. But there are all kinds of trouble getting and enforcing a patent.
errsemr
 
Posts: 23
Joined: Tue May 25, 2010 2:47 pm

Re: Encryption, the next "feature" for 2.4GHz?

Postby JiB » Tue Jun 01, 2010 8:20 pm

As Errsemr says, above, encryption may have too many downsides for the manufacturers too. The Lock-in works and is very profitable. But it would be neigh to impossible to introduce encryption, while still being compatible to your locked-in customers. Your customers may be locked-in to your system, but you are too.
Besides that, encryption is fairly pointless with these chips. The question rather becomes, can you read the flash on a microcontroller? The answer to that is a tentative yes, depending on the tools you have access too. Once the bad guys have access to the memory, they can disassemble the software, figure out the pseudo random number generator on it and decrypt the whole datastream.
Although software has no marginal cost, it adds significantly to the cost of any manufacturer. Adding encryption would only increase this cost, and will not have large enough payoff, I think. But it is all crystal ball gazing, we just have to see when we get there.
JiB
 
Posts: 11
Joined: Tue Jun 01, 2010 7:58 pm

Re: Encryption, the next "feature" for 2.4GHz?

Postby RCModelReviews » Tue Jun 01, 2010 10:03 pm

The benefit to the manufacturer isn't so much that it stops their product being cloned but that it provides very simple and irrefutable proof that a product *is* a direct copyright violation and does so in a manner that any court would recognize straight away.

If a counterfeiter simply copies the MCU then the real manufacturer has to prove that they've done so and this involves expensive reverse-engineering as well as complex technical evidence, expert witnesses and other expense if a ban is sought on the importation and sale of such devices.

When encryption is used, the mere fact that a product works is proof that it's been unlawfully copied.

There would be *no* legal "compatible" versions of a system that had encryption because in order to work, the MCU would have had to been copied.

Once a court ruling was obtained, all imports could be seized (under counterfeiting laws) and those found to be doing the importing (for personal reasons or for profit) could face stiff penalties. That would kind of kill the attractiveness of cheap rip-offs wouldn't it?
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Encryption, the next "feature" for 2.4GHz?

Postby JiB » Tue Jun 01, 2010 11:07 pm

Sorry Bruce, but proof of copyright infringement does not follow from use of encryption. Once you can disassemble the code, you are still not doing something wrong (perhaps unless you live in one of these DMCA dictatorships); it is your product, you are free to look at it any way you like. Disassembly may mean reassembly, and that would indeed be a copyright infringement. Having a look at the (assembler)code may also mean you can abstract the algorithm from it, which you can implement in your own code. That would not be a copyright violation, and when the stakes a high enough, it would be a worthwhile endeavour.
So encryption of the datastream is not an easy indicator of a copied firmware. It might be, but the claimant would still have to prove that. Since the processing power of most uC's is fairly low and all sidebands (power consumption, reaction timings) would also be available, it is quite feasible the encryption would have been brute forced.
JiB
 
Posts: 11
Joined: Tue Jun 01, 2010 7:58 pm

Re: Encryption, the next "feature" for 2.4GHz?

Postby RCModelReviews » Tue Jun 01, 2010 11:47 pm

JiB wrote:S Having a look at the (assembler)code may also mean you can abstract the algorithm from it, which you can implement in your own code. That would not be a copyright violation, and when the stakes a high enough, it would be a worthwhile endeavour.

But... you'd still need to copy the the encryption key and that would be all the proof required in most courts of law.

Even having the exact same algorithm will be useless unless you have the necessary decryption key and the only way to get that is (short of commissioning massive amounts of computing power for long periods of time) to illegally copy it.

So encryption of the datastream is not an easy indicator of a copied firmware. It might be, but the claimant would still have to prove that. Since the processing power of most uC's is fairly low and all sidebands (power consumption, reaction timings) would also be available, it is quite feasible the encryption would have been brute forced.

Providing a strong algorithm and a sufficiently sized key is used you do need a *lot* of brute-force.

Just look at what Digi are doing with the XBee modules. They have 128bit AES encryption as a standard feature and that doesn't seem to introduce any latency. Those modules are cheap-as-beans so there's not *that* much processing power required to support this kind of strong encryption.
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Encryption, the next "feature" for 2.4GHz?

Postby JiB » Wed Jun 02, 2010 2:15 pm

RCModelReviews wrote:
JiB wrote:S Having a look at the (assembler)code may also mean you can abstract the algorithm from it, which you can implement in your own code. That would not be a copyright violation, and when the stakes a high enough, it would be a worthwhile endeavour.

But... you'd still need to copy the the encryption key and that would be all the proof required in most courts of law.
That is a very narrow definition of copyright. AFAIK, reverse engineering to make your product compatible with another is explicitly allowed in the US (despite the DMCA). So your argument now hinges on the copyrightability of an encryption key. For something to be eligible for copyright, it needs to contain the creative 'freedom' of the author. So on one hand, putting any text in the encryption key would give it copyright protection. But, on the other hand, the solution to a mathematical problem (like 2+2=4) is obviously not copyrightable. In that view, the encryption key can be seen as the outcome of the mathematical problem f(cyphertext, cleartext) = key. Thus, the key is not copyrightable and you cannot misuse copyright to 'protect' your encryption. If you wanted your copyrighted text to remain protected, you should not have put it in the encryption key.
On top of that, in programming, you cannot get copyright on a specific way to do something (like an algorithm), especially if that is the only way to solve a specific problem. (note, this is copyright, not patents and I am no attorney in either/at all. Patents is a whole different minefield). By using a specific key and algorithm, you limit the options to one, you can only decrypt with that algorithm and key. Thus you have removed the part that makes something copyrightable and in turn you cannot claim copyright on the key or algorithm you used.

Besides that, it is a moot point, because my other arguments still stand: it is very detrimental for a manufacturer that employs lock-in to abandon their current system to introduce something incompatible that uses encryption. The downsides have to weigh up to the upsides, and there is still a lot of money in the upside. That might change when the cheap compatibles have eroded a lot of their profit margins, but I have not seen that happen yet: AFAIK, genuine DSM2 or FASST receivers still cost you an arm and a leg and they still have a loyal base of customers buying these overpriced products.
JiB
 
Posts: 11
Joined: Tue Jun 01, 2010 7:58 pm

Re: Encryption, the next "feature" for 2.4GHz?

Postby johnkim100530 » Sun Jun 13, 2010 9:01 pm

There is no need for cloned receivers. If you need a cheap after-market receiver, buy a matching module . Chinese modules cost only around $20.
What should I do with Futaba module and receiver I already bought? Just use them in a very expensive plane or sell them.
johnkim100530
 
Posts: 61
Joined: Sat May 08, 2010 8:23 pm
Location: Panorama City, California

Re: Encryption, the next "feature" for 2.4GHz?

Postby stoney7713 » Wed Jun 16, 2010 7:05 pm

I'm about fed up with all the different 2.4Ghz technology as is. I remember reading something about the Futaba 8 ch TX not working with with some of the lower channel RX's? You can't use Hitec or Futaba TX and RX and vise verse, that use to be the easy part of grabbing gear to go fly, piecing together your gear, reliable gear on 72Mhz.

I wish there was a industry standard, like with wireless networks, blue ray, 72Mhz TX and RX.....

I own a Hitec Optic 6 Spectra TX, I thought of converting it over to 2.4, but I didn't like the particular way of its binding feature or the search feature, I was afraid noise could still pop up and cause loss of control of my planes. So I bought a Futaba 7C, got an extra channel, and what I see as being the best at avoiding interference, as long as I don't get locked into using Futaba servos too, I'll be fairly happy with the choice I made.

Like everyone, i just hate the expensive receivers, and if someone can clone, (not copy) them and sell them for half price I'd be more then happy if they work, but for the same reason I went with with the Futaba radio, I wanted reliability, not a Chinese knock off that works 75% of the time.

I don't care if they start encrypting newer radios, that is as long as mine still works. Like with wireless networking, just because my wireless card cannot do the newest encryption technology, the wireless router is backwards compatible and I can still use the 2 together. In this case the receiver would have to be backwards compatible.

I also know there are more and more 2.4Ghz generic TX and Rx on the market, and even those you must use their TX and RX combo for them to work.

I wonder if there will ever be an industry standard for 2.4Ghz, or maybe we will have to wait for the new magic frequency.
stoney7713
 
Posts: 3
Joined: Wed Jun 16, 2010 5:49 pm

Re: Encryption, the next "feature" for 2.4GHz?

Postby RCModelReviews » Wed Jun 16, 2010 10:02 pm

stoney7713 wrote:I wanted reliability, not a Chinese knock off that works 75% of the time.

Interestingly enough, some of the newer Chinese product (FrSky for instance) has a better track record in terms of reliability than either Futaba or JR/Spektrum.

I've been flying "Chinese" product for a long time now and haven't had a single radio-related issue.

I've seen several JR/Spektrum and at least one FASST model "bite the dust" for unexplained reasons during that time.

Just as a Toyota Corolla isn't a Ferrari, so a Chinese 2.4Ghz isn't a Weatronics -- but there's absolutely nothing wrong with a good old Corolla, they are one of the most reliable vehicles on the road and very affordable.

Nothing wrong with Futaba .. except the price (and the lack of telemetry options).
RCModelReviews.com, just the facts.
User avatar
RCModelReviews
 
Posts: 2120
Joined: Tue May 04, 2010 3:40 am

Re: Encryption, the next "feature" for 2.4GHz?

Postby stoney7713 » Thu Jun 17, 2010 4:59 am

I have watched and read your reviews, and I enjoy the unbiased remarks, it's also what we need right now with all the diferent 2.4 systems coming out.

I still like to see one on the Futaba, but just random post on RCGroups.com it seems I see a lot more post on syncing issues with some of the lesser name products, and I've read of more failure rates, where people have lost total control of their model then with Futaba.

Of course that all could be operator error, who knows, we don't exactly get to see a computer print out or slow motion recordings of all these glitches.

I know Futaba has it's issues with heat and voltage, and I don't care much for the duel antenna, and price, but I guess it comes down to how much money you have and what type of bells and whistles you want and what you feel safe with.

I may switch my Hitec over to 2.4Ghz yet for some fun with the telemetry and to pass it on to my daughters.

I drive a Kia, there's good and bad things about it, but it's what I own.

I'll let everyone know if I run into any issues with the Futaba, I still wish there was an industry standard to all of this.
stoney7713
 
Posts: 3
Joined: Wed Jun 16, 2010 5:49 pm

Previous

Return to Radios and Servos

Who is online

Users browsing this forum: No registered users and 2 guests

cron